diff --git a/cogs/tts.py b/cogs/tts.py
new file mode 100644
index 0000000..ffe17b0
--- /dev/null
+++ b/cogs/tts.py
@@ -0,0 +1,108 @@
+import io
+from contextlib import closing
+
+import aioboto3
+import discord
+from discord.ext import commands
+
+from libraries import emoji_literals
+from modules import exceptions
+from modules.misobot import MisoBot
+
+
+class GenderSelect(discord.ui.Select):
+    def __init__(self, voices):
+        self.voices = voices
+        options = [discord.SelectOption(label="Male"), discord.SelectOption(label="Female")]
+        super().__init__(placeholder="Choose voice gender", options=options)
+
+    async def callback(self, interaction: discord.Interaction):
+        for child in filter(lambda c: isinstance(c, VoiceSelect), self.view.children):
+            self.view.remove_item(child)
+
+        self.view.voice_select = VoiceSelect(
+            filter(lambda v: v["Gender"] == self.values[0], self.voices)
+        )
+        self.view.add_item(self.view.voice_select)
+        await interaction.response.edit_message(view=self.view)
+
+
+class VoiceSelect(discord.ui.Select):
+    def __init__(self, voices):
+        options = [
+            discord.SelectOption(
+                label=voice["Id"],
+                description=voice["LanguageName"],
+                emoji=emoji_literals.NAME_TO_UNICODE.get(
+                    f":flag_{voice['LanguageCode'].split('-')[1].lower()}:"
+                ),
+            )
+            for voice in voices
+        ]
+
+        # The placeholder is what will be shown when no option is chosen
+        # The min and max values indicate we can only pick one of the three options
+        # The options parameter defines the dropdown options. We defined this above
+        super().__init__(
+            placeholder="Choose TTS voice",
+            options=options,
+        )
+
+    async def callback(self, interaction: discord.Interaction):
+        # Use the interaction object to send a response message containing
+        # the user's favourite colour or choice. The self object refers to the
+        # Select object, and the values attribute gets a list of the user's
+        # selected options. We only want the first one.
+        # await interaction.response.edit_message()
+        pass
+
+
+class DropdownView(discord.ui.View):
+    def __init__(self, voices):
+        super().__init__()
+        self.add_item(GenderSelect(voices))
+
+
+class TTS(commands.Cog):
+    """Text to speech"""
+
+    def __init__(self, bot):
+        self.bot: MisoBot = bot
+        self.aws_session = aioboto3.Session(
+            aws_access_key_id=self.bot.keychain.AWS_ACCESS_KEY,
+            aws_secret_access_key=self.bot.keychain.AWS_ACCESS_SECRET,
+        )
+
+    @commands.command()
+    async def tts(self, ctx: commands.Context):
+        # Create the view containing our dropdown
+        async with self.aws_session.client("polly", "us-east-1") as polly:
+            voices = await polly.describe_voices(Engine="neural")
+
+        view = DropdownView(voices["Voices"])
+        await ctx.send("**Text to speech generator**", view=view)
+        # Wait for the View to stop listening for input...
+        await view.wait()
+        if view.value is None:
+            print("Timed out...")
+        elif view.value:
+            print("Confirmed...")
+        else:
+            print("Cancelled...")
+
+    async def generate_audio_file(self, text):
+        async with self.aws_session.client("polly", "us-east-1") as polly:
+            response = await polly.synthesize_speech(
+                Text=text, OutputFormat="mp3", VoiceId="Joanna", Engine="neural"
+            )
+            if response["AudioStream"]:
+                with closing(response["AudioStream"]) as stream:
+                    data = await stream.read()
+                    file = discord.File(fp=io.BytesIO(data), filename="speech.mp3")
+                    return file
+            else:
+                raise exceptions.CommandError("Could not get AudioStream from AWS")
+
+
+async def setup(bot):
+    await bot.add_cog(TTS(bot))
diff --git a/main.py b/main.py
index c833063..9fc0587 100644
--- a/main.py
+++ b/main.py
@@ -66,6 +66,7 @@ def emit(self, record):
     # "reddit",
     "crypto",
     "kpop",
+    "tts",
 ]
 
 infrastructure_extensions = [